Two words do more commercial work in skincare than any others, and neither of them is regulated as a unit. Clinically proven is persuasive precisely because it borrows the authority of medicine while sitting inside a category that is not medicine and does not have to meet its evidential standards. The phrase is not automatically deceptive. It is almost always underspecified, and the underspecification is the point.
There is no legal definition of the phrase
There is no statutory threshold that a study must reach before a cosmetic advertiser may use the word clinical. What exists instead is a general and quite demanding principle: before making an objective claim, an advertiser must hold documentary evidence capable of substantiating it, and that evidence must support the claim as consumers will understand it.
That framework is set out in the CAP Code and is enforced by the Advertising Standards Authority. It is a good framework. Its weakness is that it operates after publication and on complaint, which means the first version of a claim reaches the market regardless.
What a cosmetic clinical study usually is
The word clinical in this context signals supervision, not medicine. A typical cosmetic study recruits volunteers, applies the product under instruction for a defined period, and takes measurements at intervals. What varies enormously is what is being measured and how.
| Type of measure | What is actually recorded | How much weight it can bear |
|---|---|---|
| Instrumental | Readings from devices measuring hydration, water loss, elasticity, gloss or roughness | The most objective of the three, though device readings are sensitive to conditions and to technique |
| Expert grading | A trained assessor scoring the skin against a defined scale, ideally blinded | Reasonable where the assessor is blinded and the scale is validated, weaker where neither is stated |
| Self assessment | Volunteers answering questions about how the product felt and looked to them | A measure of perception. Real and worth knowing, but not evidence of a physical change |
All three are frequently reported using the same vocabulary. A result described as an improvement in the appearance of fine lines might rest on a device reading, on a blinded grader, or on volunteers agreeing with a statement in a questionnaire. Those are very different claims dressed identically.
Clinically proven to reduce the appearance of fine lines
That the advertiser holds evidence from a study conducted under some form of supervision which supports a change in the appearance of fine lines. Appearance is the operative word: it is a claim about how skin looks, which can be satisfied by optical effects, hydration and light scattering as well as by any structural change.
The claim must be capable of substantiation before it is published, and the evidence must match the claim as a reasonable consumer would read it.
It does not mean a change in skin structure. It does not mean a peer reviewed publication exists. It does not imply a control group, blinding, randomisation, or any particular number of participants. It does not mean the effect was large, lasting, or greater than a plain moisturiser would produce. It does not mean the study was independent of the brand.
A version of this claim you could actually assess would state the design, the number of participants, the duration, what was measured and how, whether there was a control or comparator, and where the study can be read. Some brands publish this. Where they do not, the phrase is asserting the existence of evidence rather than presenting it.
The five questions that separate evidence from decoration
You do not need a background in statistics to interrogate a cosmetic claim. Five questions do most of the work, and the answers are either published or they are not.
- How many people. Small panels produce noisy results. A number given for a percentage without a denominator is the most common evasion in the category.
- Compared with what. Against nothing, against a vehicle without the active, or against another product. Against nothing is the weakest and the most common.
- Measured how. Instrument, blinded grader, or the volunteers' own opinion.
- Over how long. Four weeks and twelve weeks are not comparable, and neither is a single application.
- Who ran it and who paid. Brand funded research is normal across many industries and is not disqualifying, but it is material information.
A claim that answers all five is doing something real. A claim that answers none is a design element.
The percentage that is really a questionnaire
The most widespread construction in the category is a percentage of participants agreeing with a statement. It is usually accurate, it is usually properly footnoted somewhere, and it is usually read by shoppers as something it is not.
92% of users saw smoother skin
That in a consumer study of some size, that proportion of participants agreed with a statement about smoother skin. Consumer perception studies are legitimate research and the figure is usually reported accurately.
Where the claim is presented as a consumer opinion and the basis is made clear, it is a claim about what users thought.
It does not mean skin was measured. It does not mean an assessor observed a change. It does not mean the change exceeded what any moisturiser would produce. It does not tell you how many people were asked, how they were recruited, whether they knew what they were testing, or how the question was worded. A leading question and a small panel can produce a very high number without anything having happened to anyone's skin.
The footnote would have to state the number of participants, the duration, the exact statement they were responding to and the recruitment method. Where those appear in small print, read them: they usually tell you the claim is about perception, at which point the headline number is far less impressive and considerably more informative.
Where a cosmetic claim becomes a medicinal one
There is a hard boundary underneath all of this. A product presented as treating or preventing disease, or as restoring, correcting or modifying physiological functions by pharmacological, immunological or metabolic action, is a medicinal product and requires a marketing authorisation. Cosmetic claims have to stay on the correct side of that line, which is why the category is so full of words like appearance, look and feel.
The line is policed by the MHRA, whose guidance on determining whether a product is a medicinal product is the starting point for anyone trying to understand why claims are worded the way they are. It is covered in detail in the line between a cosmetic and a medicine.
Evidence as marketing asset
A study is expensive, and once it exists it has to earn its cost. The pressure is therefore not to find out whether the product works, but to generate a defensible sentence. Study design responds to that pressure: shorter durations, perception measures, no comparator, and a population selected to show the effect most clearly.
None of this requires anyone to falsify anything. Every individual choice can be defended, and the aggregate is a body of research designed to produce claims rather than knowledge.
- Testing houses. Sell studies scoped to produce a usable claim within a budget and a timetable
- Brands. Acquire a defensible sentence that competitors without a budget cannot match
- Retailers and platforms. Get substantiated copy that reduces their own compliance exposure
- Consumers. Get accurate sentences that support inaccurate impressions
A description of incentives that operate across the sector. It is not an allegation about any specific study or company.
What good evidence looks like when it appears
It does exist, and it is worth recognising so that the scepticism stays calibrated. Strong cosmetic evidence tends to share features: a published or fully described protocol, a vehicle controlled design where the comparison is the same formula without the active, blinded assessment, a duration long enough for the claimed change to be plausible, a stated participant number, and results reported with variability rather than as a single flattering figure.
Where the underlying ingredient has independent literature, you can go further and read it yourself through PubMed or the Cochrane Library. The important discipline there is to check that the concentration, vehicle and duration in the literature resemble the product in front of you, because a well evidenced ingredient in a poorly built product is a very common situation.
What to do with a clinical claim at the shelf
Treat the phrase as an invitation to look for a footnote rather than as a conclusion. If the footnote describes a consumer perception study, you have learned that people liked it, which is genuinely useful information about whether you will keep using it. If the footnote describes an instrumental study with a comparator, you have learned considerably more. If there is no footnote at all, you have learned that the brand chose not to say.
And if a claim seems to promise something a cosmetic is not permitted to do, that is not a sign of a breakthrough. It is a sign of a claim that may be on the wrong side of a regulatory line, which is a reason for caution rather than excitement. Complaints about advertising can be made to the Advertising Standards Authority, whose rulings are published and searchable.